[WFLY-19790] batch-processing: replaces -ds.xml deprecated filed with Jakarta's DataSou…

[emmartins]

…rceDefinition
Issue: https://issues.redhat.com/browse/WFLY-20905

[emmartins]

@tadamski I am seeing the following warning when using the DataSourceDefinition instead of the -ds.xml

20:12:46,089 WARN [org.jboss.jca.core.connectionmanager.pool.strategy.OnePool] (ServerService Thread Pool -- 46) IJ000407: No lazy enlistment available for batch-processing_batch-processing_batch-processing_java:jboss/datasources/batch-processingDS

Perhaps there is a property we need to add to the DataSourceDefinition to prevent this?

[emmartins]

@jfdenise FYI it seems Glow is not aware of DataSourceDefinition annotation and reports the following error in this project, if you remove the failOnError

[ERROR] * unbound datasources error: java:jboss/datasources/batch-processingDS

[emmartins]

@liweinan IMHO this is still a work in progress, after all it currently replaces one warn with another, and then there is the provisioning error... I suspect we will need server side and glow side changes to make it "proper"!

@kstekovi WDYT?

[liweinan]

@emmartins cool!

[emmartins]

@tadamski can you please have a look at #973 (comment) above once you have a chance?

[emmartins]

@kstekovi for the record I don't think we should merge this till we have that new warning suppressed or documented, we should not replace one warning with another.

[jfdenise]

@emmartins @liweinan ,
DataSourceDefinition is going to be natively supported by WildFly Glow. We have the following PR opened:

• Generic Annotation value handling and Support for DataSourceDefinition wildfly-glow#94
• Fix for WFLY-19989, h2-driver layer rule for jakarta DataSourceDefinition annotation wildfly#18420
  Once merged this would require a WildFly Maven Plugin upgrade and usage of WildFly SNAPSHOT (until WF 35.Beta1 is released).

You would then have to remove the explicit h2-datasource add-on, and the fail-on-error.

[emmartins]

thanks @jfdenise , I will keep this open till that happens.

[emmartins]

@tadamski hello?

[jfdenise]

@emmartins , FYI, The changes in WildFly have been merged. We are waiting for a new WildFly Maven Plugin release that integrates the support.

[emmartins]

@tadamski please please? ^^

[tadamski]

@emmartins @darranl sorry for this huge lag!

In order to get rid of the warning cached connection manager has to be turned on, but I don't see such possibility for annotated datasource. I'm going to create a ironjacamar fix for that but I think this PR can be merged.

FYI I have downgraded logging to DEBUG (JBJCA-1513). CCM is problematic in itself and requires refactor (see JBJCA-1494). It is perfectly fine that it is off by default in datasource annotation parsing and shouldn't introduce any cofusing warn messages.

[emmartins]

@tadamski thanks for your feedback, personally I see no interest in making this change till the warning is gone, after all this would be replacing one warning (that is already documented) with another.

[tadamski]

@emmartins This warning is only informative, tbh I was considering removing it, but after it was downgraded to debug it won't be visible anyway in standard configuration

[emmartins]

@tadamski does this means that WFLY 36 will no longer produce this warning?

[tadamski]

@emmartins yes